// the find
fuzhengwei/chatbot-api
《ChatGPT AI 问答助手》 开源免费项目，涵盖爬虫接口、ChatGPT API对接、DDD架构设计、镜像打包、Docker容器部署，小巧精悍，流程全面。对于Java编程伙伴来说，非常具有学习价值。
A small Spring Boot app that scrapes unanswered questions from the author's paid Zhishi Xingqiu (知识星球) community, sends them to the OpenAI API, and posts the generated answers back on a schedule. It's built and sold as a teaching project walking Java learners through DDD-style module layering, not as a general-purpose chatbot.
The module split (application/common/domain/infrastructure/interfaces) is a clean, legible example of layered architecture for people who haven't built one before. The whole pipeline — scrape, call OpenAI, cron job, Docker build — is small enough to read end to end in an afternoon, which is the point. Dockerfile and build.sh are included so you can see the actual deploy story, not just source.
Its only real data source is a paid Chinese platform (知识星球) accessed through what's effectively a reverse-engineered scraper — there's nothing to adapt here if you don't use that platform, since the domain logic is specific to its API shape. The README is mostly a funnel: bio, WeChat QR codes, links to the paid community and the video course, with actual setup squeezed into a couple of short notes. There's a leftover TensorFlow/Python install section with zero connection to the chatbot code, which suggests the README is copy-pasted across the author's repos rather than maintained per-project. Test coverage is a single empty-looking SpringBootRunTest/ApiTest stub, and the repo has had no commits since October 2023.